In this study, we analyzed the responses to 9 physical and 12 psychological items on the Menopausal Health-Related Quality of Life questionnaire by 491 Japanese women attending our menopause clinic aged 40-64 years. The first analysis extracted 3 PCs (depression, somatic, and vasomotor/sleep) that defined the variance of physical and psychological symptom profiles of the women. A subsequent cluster analysis was performed based on the 3 PCs to generate 4 clusters: CL8, CL6, CL5, and CL4. CL8 was composed of women who had only mild-to-moderate musculoskeletal pains and tiredness. All women in CL6, CL5, and CL4 described their musculoskeletal pains and tiredness as moderate to severe. The women in CL5 also had moderate-to-severe vasomotor symptoms, while the women in CL4 also suffered from moderate-to-severe psychological symptoms, such as depression, anxiety, and insomnia.
Many previous studies have categorized menopausal symptoms of non-menopause clinic samples and population based samples, and identified subgroups of women in the menopausal transition on the basis of their responses to symptom checklists. In 1986, Hunter et al. first introduced their “Women’s Health Questionnaire”, and they performed a PCA of responses to the 36-item symptom checklist among 682 English women attending an ovarian screening program aged 45-65 years [1]. Based on the analysis, they defined 9 symptom clusters, namely “somatic symptoms”, “depressed mood”, “cognitive difficulties”, “anxiety/fears”, “sexual functioning”, “vasomotor symptoms”, “sleep problems”, “menstrual”, and “attractive”.
Similar studies using PCA methods have been performed to identify menopausal symptom clusters specific to Hong Kong Chinese [2] and Taiwanese [3]; to examine cross-cultural variation among the United States, Spain, Lebanon, and Morocco [4]; to investigate how symptoms cluster across menopausal transition stages [5]; and to determine which symptom cluster has the greatest effect on quality of life [6]. In a random telephone survey of Hong Kong Chinese women, Ho et al. identified 5 symptom clusters: psychological, musculoskeletal/gastrointestinal, non-specific somatic, respiratory, and vasomotor [2]; in a resident cohort study of Taiwanese women, Fuh et al. identified 4 symptom clusters: musculoskeletal, non-specific somatic complaints, urogenital, and vasomotor [3]. Sievert et al. showed in their study of general population that the inter-correlation among symptoms differed in country-specific ways; for example, hot flush grouped with vaginal dryness and sexual symptoms in Spain, with general somatic symptoms in Morocco, and did not cluster with other symptoms in the United States or Lebanon [4]. Based on the observations from the Seattle Midlife Women’s Health Study (SMWHS), Cray et al. revealed similar factor structures across the 4 menopausal transition stages (late reproductive, early or late menopausal transition, or early postmenopause) in that each stage revealed a mood component, a vasomotor component, and a pain component [5]. A study by Greenblum et al. identified 3 symptom clusters in middle-aged Floridian community-dwelling women: (1) anxiety, irritability, and fatigue; (2) weight gain and urinary stress incontinence; and (3) vaginal dryness and sleep disturbances, which had the greatest impact on quality of life [6].
A different approach using cluster analysis was taken to investigate the relationships between patterns of depressed mood and menopausal transition stages [19] or to identify clusters of female urological symptoms [20]. Using the data from SMWHS, Woods et al. revealed that the majority of women experienced the menopausal transition without a high severity of depressed mood, while a small group of women had mood worsening over time and others improved [19]. Hall et al. conducted a cluster analysis of urological symptoms among women in the Boston Area Community Health Survey and identified 4 clusters, which were distinguished by the severity of storage symptoms, frequency symptoms, urinary incontinence, and interference [20].
Recently, Cray et al. introduced latent class analysis to identify subgroups of women in the menopausal transition stage that experienced the same cluster of representative symptoms [7],[8]. They first selected 5 symptoms based on PCA of the SMWHS participants’ responses to the 47-item symptom checklist: sleep, pain, mood, cognitive, and tension. Multilevel latent class analysis using scores for hot flush and these 5 symptoms identified 3 classes: low severity levels for all symptoms; low severity hot flush and moderate severity levels for all other symptom factors; high severity hot flush with lower severity levels of all other symptom factors. In their most recent report, Woods et al. identified several endocrine biomarkers associated with the distinction among the subgroups, including estrogen, follicle stimulating hormone, epinephrine, and norepinephrine [21].
Similar to our findings, a 2013 SMWHS report by Cray et al. identified 3 major components of symptoms identified by PCA [5]. These included psychological, vasomotor, and musculoskeletal pains, of which the psychological component accounted for more than 30% of variance of the profile of menopausal symptoms. Cross-cultural diversity in menopausal symptomatology has been frequently demonstrated [22], and fewer Asian women have been shown to report vasomotor and other menopausal symptoms compared with their Caucasoid counterparts [23]; nevertheless, these fundamental components might be similar between the United States and Japan.
One of the major limitations of our study was that the participants were recruited in a clinical setting, and the results were not necessarily representative of the general population of Japan. Nevertheless, the findings of the current study are relevant to clinical practice as they address the complex symptomatology in women attending a menopause clinic.
